Serious congenital heart malformations are present in approximately one percent of all live births in humans. Studying the pathogenesis of congenital heart malformations can lead to more accurate expectations of associated pathologic lesions, and thus to more rational management of affected patients. The present study is an interdisciplinary effort which involves collection and computer processing of descriptive and quantitative data from staged normal human embryos in the Carnegie Embryological Collection and from patients autopsied at The Johns Hopkins Hospital. These data and proposed hypotheses of pathogenesis are translated into symbolic logic, which is then used to determine all and only the conclusions which are logically deducible from the data and hypotheses. Direct observations, non-controversial concepts, and speculative hypotheses are stratified in descending order of certainty, and the computer program always deduces a conclusion based upon more certain information in preference to one based upon less certain information. This method has now been used to identify pathologic lesions associated with tetralogy of Fallot, malformations of the semilunar valves, atrioventricular canal malformations, interrupted aortic arch, double inlet left ventricle, and corrected transposition which are not well explained by existing theories of pathogenesis. Similar investigations are currently in progress for truncus malformation, ventricular septal defect, double outlet right ventricle, and persistent left superior vena cava. Discrepancies, when identified, form the basis for modification of the pathogenetic hypotheses. The ultimate product of this research will be pathogenetic hypotheses which are stated in the objective language of symbolic logic, and which account for the observed embryologic and autopsy data with statistically sound explanations.